Insights into population structure and epidemiology of Phytophthora infestans from Nicaragua

(1)

Insights into population structure and epidemiology of Phytophthora

infestans from Nicaragua

Jorge Ulises Blandón-Díaz

Faculty of Natural Resources and Agricultural Sciences Department of Forest Mycology and Pathology

Uppsala

Doctoral Thesis

Swedish University of Agricultural Sciences

Uppsala 2011

(2)

Acta Universitatis agriculturae Sueciae

2011:36

ISSN 1652-6880

ISBN 978-91-576-7581-1

Cover: Potato and tomato leaves affected by the late blight pathogen Phytophthora infestans (Mont.) de Bary.

(photo: Jorge Ulises Blandón-Díaz)

(3)

Insights into population structure and epidemiology of Phytophthora infestans from Nicaragua.

Abstract

Late blight caused by Phytophthora infestans (Mont) de Bary is a constraint to both potato and tomato crops in the northern highlands of Nicaragua. This thesis describes studies on population structure and epidemiology of P. infestans from Nicaragua.

The genotypic and phenotypic variation in isolates of P. infestans collected in potato and tomato growing areas of northern Nicaragua were analyzed using genotypic (SSR fingerprinting and mtDNA haplotyping) and phenotypic markers (mating type, virulence and fungicide sensitivity). Genotypic markers revealed no polymorphism among the P. infestans isolates tested. Phenotypic variation was observed. Nicaraguan population of P. infestans is dominated by a clonal lineage of the A2 mating type, Ia mtDNA haplotype and no evidence of genetic population differentiation among potato and tomato isolates was found.

The aggressiveness of P. infestans isolates sampled from potato and tomato fields was determined through cross-inoculations experiments. Potato and tomato isolates both had a shorter LP, higher SP, and were more aggressive on tomato leaflets compared to potato ones.

The adequacy of the late blight simulation model LATEBLIGHT (version LB2004) was evaluated under Nicaraguan conditions. During 2007-2008 field experiments were conducted in Nicaragua. The simulation model was considered adequate as it accurately predicted high disease severity in susceptible cultivars without fungicide sprays, and demonstrated a decrease in the disease progress curves with additional fungicide applications, similar to that observed in the field plots. The quantitative relationship between host resistance and the need for fungicide was also investigated using simulations performed with LATEBLIGHT, as well as field trials.

Keywords: pathogen diversity, quantitative pathogenicity, epidemiology, modelling, host resistance, Phytophthora infestans, fungicide resistance

Author’s current address: Jorge Ulises Blandón-Díaz, SLU, Department of Forest Mycology and Pathology, P.O. Box 7026, SE-750 07 Uppsala, Sweden E-mail: ulises.diaz.blandon@slu.se

Author’s home address: Departmento de Protección Agrícola y Forestal, Universidad Nacional Agraria, km 12 Carretera Norte, Apdo 453, Managua, Nicaragua. E-mail:

ulises.diaz.blandon@una.edu.ni

(4)

(5)

Dedicada

A Jürgen, Jorge Ulises Jr., Lorena y a mi querida familia en Estelí y Managua.

(6)

List of Publications

This thesis is based on the work contained in the following papers, referred to by Roman numerals in the text:

I Blandón-Díaz, J.U., Widmark, A.-K., Hannukkala, A., Andersson, B., Högberg, N., and Yuen, J.E. (2011). Phenotypic variation within a clonal lineage of Phytophthora infestans infecting both tomato and potato in Nicaragua. Phytopathology. (Accepted with revisions).

II Blandón-Díaz, J.U., Högberg, N., Grönberg, L., Widmark, A.-K., and Yuen, J.E. (2011). Aggressiveness and genotyping of Phytophthora infestans isolates from Nicaragua. (Manuscript).

III Blandón-Díaz, J.U., Forbes, G.A., Andrade-Piedra, J.L., and Yuen, J.E.

(2011). Assessing the adequacy of the simulation model LATEBLIGHT under Nicaraguan conditions. Plant Disease. (In press).

IV Blandón-Díaz, J.U., Forbes, G.A., Taipe, A., Knutsson, J., Andrade- Piedra, J.L., and Yuen, J.E. Epidemiological significance of the quantitative relationship between host resistance and fungicide usage.

(Manuscript).

Paper III is reproduced with the permission of the publishers.

(10)

(11)

Abbreviations

AI Aggressiveness index ANOVA Analysis of variance

AUDPC Area under disease progress curve CTAB Cetyltrimetylammonium bromide DNA Deoxyribonucleic acid

EAT Envelope of acceptance test GLM Generalized linear model

H_hr Hours per day of relative humidity >85%

IP Incubation period LA Lesion area

LB2004 Lateblight model version 2004 LGR Lesion growth rate

LP Latency period LSMEANS Least-squares means mtDNA Mitochondrial DNA PCR Polymerase chain reaction

RAUDPC Relative area under disease progress curve SA Sporulating area

SCRI Scotish Crop Research Institute SP Spore production

SR Sporulation rate

SSR Simple sequence repeats

(12)

(13)

1 Introduction

In Nicaragua potato production is concentrated in three northern departments (provinces), namely, Estelí, Jinotega and Matagalpa. The production is dependent on imported seed and is expensive. The potato is grown on hillside lands positioned at altitudes ranging from 700 to 1500 meters above sea level (masl), although it has also been grown in the Sebaco valley located at 400 masl. Under optimal growing conditions, the yields of potato range from 20 to 25 t ha^-1. Potato consumption in Nicaragua is about 2727.3 tons per month, except December when it increases to 4090.9 tons. The national production does not meet domestic demand even though the country has potential planting areas (355,233 ha), which could be used for potato production (^PFID-^F&V, 2005). In 2008 and 2009, Nicaragua produced 33000 metric tons in 2300 ha (FAOSTAT, 2008). The average annual import volume represents 55% of apparent consumption, while domestic production accounts for 45% (PFID-F&V, 2005). As in many countries around the world, the potato crop in Nicaragua is attacked by many pests and diseases that require producers to use large amounts of insecticides, fungicides and bactericides. One of the most important diseases affecting potato crop in Nicaragua is late blight, caused by the oomycete Phytophthora infestans. This pathogen also affects tomato, which is grown year round in altitudes ranging from 400 to 1500 masl. The area allocated for the tomato crop is approximately 2000 ha and the yields range from 12 to 18 t ha^-1 (Berlin and Eitren, 2005). Potato and tomato are grown in adjacent areas increasing the risk of passing the late blight from one crop to another if one of them becomes infected. However, these crops rarely are grown in the same field.

This thesis describes aspects until recently unknown about P. infestans in Nicaragua. This oomycete causes late blight disease on potato and tomato

(14)

crops. The genus name (Phytophthora) in Greek means literally and accurately the “plant destroyer”.

Sampling of potato and tomato fields affected by late blight in northern Nicaragua was carried out from 2007 to 2010. From these samples, axenic cultures of P. infestans were obtained and used for genotypic and phenotypic characterization of the plant pathogen. Genotypic characterization was done using simple sequence repeats (SSRs, also known as microsatellites) and mitochondrial DNA (mtDNA) haplotyping. The pathogen was phenotypically characterized through mating type determination, virulence and fungicide sensitivity testing. Aditionally, cross-inoculation experiments were performed to determine the level of variation of aggressiveness in P.

infestans isolates sampled from potato and tomato fields. A slight genotypic variation among the tested P. infestans isolates was found. Nonetheless, phenotypic variation (high levels of metalaxyl resistance and race diversity) among tested isolates was shown. Moreover, there seems to be some kind of specialization toward tomato based on the aggressiveness tests (Paper I and II).

Field experiments during 2007-2008 in two northern regions (Estelí and Matagalpa) of Nicaragua were set up to assess the adequacy of the late blight simulation model LATEBLIGHT (version ^LB2004) under Nicaraguan conditions. Two susceptible (Cal White and Granola) and one resistant (Jacqueline Lee) potato cultivars were evaluated, without use of fungicide, and with three application intervals (4, 7 and 14 days) of the fungicide chlorothalonil. The simulation model was considered adequate as it accurately predicted high disease severity in susceptible cultivars without fungicide protection, and demonstrated a decrease in the disease progress curves with additional fungicide applications, similar to that observed in the field plots. The model also generally predicted inadequate fungicide control, even with a 4-day spray interval, which also occurred in the field. Lack of adequate fungicide protection would indicate the need for cultivars with higher levels of durable resistance, and that farmers should consider more effective fungicide applications (higher dosages or different chemistries) if susceptible cultivars are used. The LATEBLIGHT model was also used to investigate the quantitative relationship between host resistance and fungicide usage (Paper III and IV).

(15)

2 Background

2.1 An overview of population biology/structure of Phytophthora infestans (Mont.) de Bary

In agricultural systems, plants are constantly exposed to a wide range of pathogenic microorganisms, including bacteria, fungi and oomycetes. Fungi and oomycete plant pathogens cause many of the world’s most notorious plant diseases, which threaten global food production and consequently food security. Fungi and oomycetes are in general divided into biotrophs, necrotrophs and hemibiotrophs, based on the different strategies they use to colonize plants. Biotrophic plant pathogens establish a nutritional relationship with living host cells; necrotrophic pathogens rapidly kill host plant tissues; and hemibiotrophic plant pathogens have both a biotrophic and necrotrophic phase during its life cycle (Bouwmeester et al., 2009;

Dodds and Rathjen, 2010; Schneider and Collmer, 2010).

More than 150 years after the Irish potato famine, Phytophthora infestans, the causal agent of late blight disease, continues to be an economically important pathogen of potato and tomato worldwide (Ristaino, 2002). The disease was reported first in 1843 in some northeastern areas of United States, from where it spread to others parts of the country and Canada. In Europe, the disease was reported in Belgium, Holland, Germany, Switzerland, France and Italy in the middle of 1845, spreading rapidly to England, Scotland and Ireland in the same year. In the latter country the pathogen found the optimal environmental conditions for its development and caused devastating epidemics leading to the infamous great potato famine by mid-October 1845, which resulted in ecological and social disaster in Ireland (Peterson, 1992; Ristaino, 2002; Scholthof, 2007). After

(16)

the Irish potato famine, the pioneering work of early mycologists in the identification of P. infestans and further elucidation that it was the cause and not the effect of the disease known as late blight laid the foundations for the disciplines of Microbiology and Plant Pathology (Judelson and Blanco, 2005; Ristaino, 2002).

The late blight pathogen, P. infestans, is a diploid, heterothallic and hemibiotrophic oomycete that poses a real and potential threat not only to economically important crops such as potatoes and tomatoes, but also for tree and shrub species of the family Solanaceae (Bouwmeester et al., 2009;

Fry, 2008). The “devastating plant destroyer”, P. infestans, causes economic losses yearly calculated in multibillion dollars (Haverkort et al., 2008). P.

infestans can reproduce asexually and sexually. Sexual reproduction in this heterothallic oomycete only occurs when two mating types termed the A1

and A2 outcross. As a result of this, oospores are produced, which can survive in the absence of a host (Drenth et al., 1995; Ristaino, 2002). In locations with only the asexual cycle, P. infestans survives as mycelium in infected potato tubers and debris (Ristaino, 2002) and probably also in alternate wild hosts. As a hemibiotrophic plant pathogen, P. infestans has both a biotrophic and necrotrophic phase during its life cycle (Bouwmeester et al., 2009). In compatible interactions with potato, the biotrophic phase of P. infestans can last from three to five days, after which macroscopic symptoms are evident (necrotrophic phase). In tomato leaves, an extended period of biotrophy has been observed due to a compatible interaction with tomato-specialized isolates (Legard et al., 1995; Smart et al., 2003; Vega- Sánchez et al., 2000).

Prior to the 1980s, worldwide populations of P. infestans were dominated by a single clonal lineage known as the US-1 “old” genotype, with the A1 mating type (Fry and Goodwin, 1997b; Goodwin et al., 1994a).

In contrast, in the Toluca Valley in central Mexico the ^A1 and ^A2 mating types were present in approximately equal frequencies and the populations of P. infestans were entirely different from populations in other locations (Fry, 2008; Goodwin et al., 1992). Since the mid-1980s changes in the population structure of P. infestans outside Mexico have been reported (Fry and Goodwin, 1997a). These changes brought about the displacement of the ‘old’ genotypes by ‘new’ ones, which are characterized by increased fitness and aggressiveness in addition to metalaxyl resistance (Day and Shattock, 1997). The pathogen has been found to be reproducing sexually under field conditions outside its putative center of origin. There are also reports of oospore formation and oospores acting as initial inoculum (Andersson et al., 1998; Lehtinen and Hannukkala, 2004). High population

(17)

diversity has been found worldwide in studies conducted using molecular markers (Forbes et al., 1998). The appearance of fitter and more aggressive strains has prompted the development of methods for rapid detection and identification of these strains in order to design and implement control measures (Trout et al., 1997). The development of high-throughput codominant markers can undoubtedly contribute to the understanding of P.

infestans population biology, epidemiology, ecology, genetics and evolution as a prerequisite for devising appropriate management practices (Cooke and Lees, 2004).

Populations of P. infestans have been characterized using a series of genotypic and phenotypic markers. Phenotypically, populations of P.

infestans have been distinguished through determination of the mating type, virulence spectrum and metalaxyl resistance (Fry et al., 1993). The genotypic characterization of P. infestans strains has included the use of allozyme patterns, mitochondrial DNA (mtDNA) haplotype determination,

AFLP and RFLP fingerprints with the probe RG57. In a recent study, results from mtDNA haplotyping and RFLP analysis led to the suggestion that multiple migrations of P. infestans into China have occurred (Guo et al., 2010). In comparison with the previously mentioned markers, simple sequence repeat markers (SSRs, also referred to as microsatellites) seem to offer the greatest potential across a wide range of applications (Cooke and Lees, 2004). Over the past ten years, SSRs have been developed for the study of P. infestans (Knapova et al., 2001; Knapova and Gisi, 2002; Lees et al., 2006). Microsatellite markers have been used recently to infer that multiple introduction events of P. infestans have taken place in France and that P. infestans populations from this country are composed by two differentiated genetic cluster of isolates (Montarry et al., 2010).

In Latin America, P. infestans populations have been extensively studied.

In the Toluca Valley in central Mexico P. infestans reproduces sexually and the two mating types (A1 and A2) are found in approximately equal frequencies. In other countries of the subcontinent, P. infestans appears to reproduce primarily asexually, although both mating types have been found in the same host. For instance, initial studies in Ecuador reported the presence of two clonal lineages (EC-1 and US-1) of the A1 mating type (Forbes et al., 1997). However, further studies revealed the occurrence of the two mating types (A1 and A2) of P. infestans sensu lato in the same host (Solanum muricatum) (Adler et al., 2002), but the A2 isolate in that study is probably Phytophthora andina (Oliva et al., 2010); in Peru an ^A1 clonal population is reported (Garry et al., 2005; Pérez et al., 2001); in Brazil, the

A1 and A2 mating types are found in tomato and potato respectively, but

(18)

they have never been found in the same field (Reis et al., 2003). In Uruguay, only the A2 mating type has been found (Deahl et al., 2003); in Colombia, the Â1 and Â2 mating types have been found in the same host, Physalis peruviana (cape gooseberry), although no evidence of sexual recombination has been reported so far (Vargas et al., 2009); in Costa Rica, a clonal lineage of Â1 mating type has been reported attacking potato, however, isolates of the A2 mating type has been also found in wild Solanum species (Gómez-Alpizar, 2004). In Argentina, P. infestans populations seem to be more diverse when compared with other Latin American countries since both A1 and A2 mating types have been found. However, Argentinean isolates of the A2 mating type have been found in a higher frequency and they showed greater aggressiveness and an increased resistance to metalaxyl (Andreu et al., 2010). Clonality of P. infestans populations has been also reported from Venezuela, where only the A1 mating type has been reported (Briceño et al., 2009).

A persistent problem in plant-pathogen interactions is poor understanding of the basis of host specificity, that is, what factors determine the taxonomic range of hosts that can be infected by a specific plant pathogen. This is a fundamental question that relates both to the co- evolution of host susceptibility and pathogen virulence. In general, there is very limited knowledge about the genetics and mechanisms involved in host specificity, although host specificity varies among plant pathogens and could be determined by the phylogenetic distance between plants. The host range may include a large number of plant species at one extreme or only a single genotype of a single plant species at the other (Barret et al., 2009;

Gilbert and Webb, 2007). The effects inflicted by plant pathogens on their host plants may be qualitative and quantitative and, hence, there are qualitative and quantitative specificity. Qualitative specificity does not allow a particular plant pathogen to infect many hosts, whereas, when the specificity is quantitative, the plant pathogens have a lower performance on certain host plants (Barret et al., 2009).

The host range of P. infestans, generally, has been considered to be restricted to two important crops, potato and tomato, various wild species in the genus Solanum and also some nonsolanaceous species (Adler et al., 2002; Adler et al., 2004; Erwin and Ribeiro, 1996; Kroon, 2010).

Nonetheless, the factors determining this host range remain unknown (Kamoun and Smart, 2005). Some degree of pathogenic specialization of P.

infestans to potato or tomato has been reported (Fry, 2008; Legard et al., 1995; Oyarzun et al., 1998; Suassuna et al., 2004; Vega-Sánchez et al., 2000; Wangsomboondee et al., 2002). In USA, for example, the US-8

(19)

genotype has been detected to occur on potatoes and the US-7 and US-17

genotypes have been recovered from tomatoes (Goodwin et al., 1998). A study in Kenya and Uganda clearly showed that late blight epidemics in potato and tomato were caused by two separate, host-adapted populations of P. infestans (Vega-Sánchez et al., 2000). In contrast to this, a recent study carried out in Taiwan showed no host specificity on potato or tomato among P. infestans isolates from tomato (Chen et al., 2008). Negative relationships (genetic trade-offs) between qualitative and quantitative traits required to infect one or another host can drive the appearance of pathogenic specialization mediated by antagonistic pleiotropy, in which one or more genes favour pahtogen’s performance in one host, but impair its performance in another (Barret et al., 2009; Kawecki, 1998; Pariaud et al., 2009).

Migration events have contributed in shaping the population structure of P. infestans in a number of locations around the world and Nicaragua does not appear to be the exception. Deductively, it may be possible to formulate three hypotheses to explain the presence of P. infestans on potato and tomato fields in Nicaragua. Two of these hypotheses are related to human-mediated migration events; while the third hypothesis is that P.

infestans has always been present in Nicaragua. Firstly, it is believed that cultivated potato (Solanum tuberosum) was introduced to Nicaragua in the early 1900’s, and with it the late blight pathogen, and since then they have coevolved. Troops from United States supposedly brought potato tubers for consumption, some of which fell into the hands of local people who began growing potatoes. Therefore it could be hypothesized that the first populations of P. infestans present in Nicaragua belonged to the single “old”

clonal genotype (US-1 genotype) of the A1 mating type and the Ib mtDNA

haplotype (Fry and Goodwin, 1997a). Moreover, this introduction could have occurred before the second worldwide migration of P. infestans that has been suggested to have taken place in the 1980´s. This migration spread new genotypes that were, resistant to phenylamide fungicides and showed an increased aggressiveness to potato (Goodwin et al., 1994b). Secondly, infected potatoes or tomatoes with P. infestans could have been carried from northwestern Mexico to Nicaragua in the 1950s. This hypothesis is based on the fact that the mitochondrial haplotypes Ia and IIb have been found in herbarium specimens dating from 1954 and 1956 respectively (May and Ristaino, 2004). The possible route of migration events that led P. infestans to Nicaragua is shown in Figure 1. Thirdly, another possibility to explain the presence of P. infestans in Nicaraguan potato and tomato fields is that native plants of the Solanaceae family could have hosted a pre-existing

(20)

population of P. infestans long before the introduction of the potato as a crop. Twenty-two genera and 118 species of the Solanaceae family are found in Nicaragua. The genus Solanum alone contains 45% of the total number of species in this family in Nicaragua (Stevens et al., 2001). The pepper (Capsicum annuum) might be the first domesticated plant that served as host to P. infestans, since it was grown by the natives when the Spanish conquerors arrived in Nicaragua in the early sixteenth century (www.wikipedia.org).

(21)

US-1 genotype 1912-1933?

US-6 genotype 1950s?

Figure 1. Possible route of P. infestans migration to Nicaragua from United States in the early 1900s (US-1 genotype) and from northwestern Mexico in the middle of the 1950s (US-6 genotype).

(22)

2.2 Management of late blight disease

Suggested management strategies for late blight include use of clean seed, elimination of real and potential sources of inoculum (infected cull piles, volunteer potato plants, and wild alternate hosts), fungicides, decision support systems (DSS), intercropping, cultivar mixtures and extended crop rotations (3-4 years) to avoid early infections developed from oospores.

Along with these strategies, the use of resistant cultivars against late blight is of utmost importance, especially in locations where environmental conditions are conducive for disease development and potato growers cannot afford the numerous fungicides required to control the disease (Andersson et al., 2009; Fry, 2008; Kirk et al., 2005; Lehtinen et al., 2009;

Pilet et al., 2006). Potato growers have relied primarily on fungicide use for late blight control, both in developed and developing countries (Andrivon et al., 2006). However, environmental and health concerns accompanying pesticide use (Shtienberg et al., 1989) combined with the fungicide application costs, has prompted the search for more economically and environmentally sound control measures. Currently, strategies for late blight management aim to reduce the population size and growth rate of the pathogen in order to delay the epidemic onset and subsequently to reduce disease severity (Li et al., 2009).

Three types of resistance against potato late blight have been identified:

race-specific resistance (RS), race-nonspecific resistance (RNS) associated with late maturity and race-nonspecific resistance (RNS) which acts irrespective of maturity (Vanderplank, 1957). Race-specific resistance act delaying the epidemic onset without altering the apparent infection rates, whereas race-nonspecific resistance decreases the speed of epidemic progress without influencing the date of initial disease outbreak (Parlevliet, 1979;

Vanderplank, 1968). Moreover, race-specific resistance is thought to provide complete protection, but only to certain races of the pathogen species and is governed by a single gene or a small number of related genes.

This type of resistance is considered less durable. The race-nonspecific resistance involves multiple genes, provides only partial, but more lasting protection (McDonald and Linde, 2002). For many years, potato breeders have attempted to defeat P. infestans and its arsenal of pathogenicity factors (effectors) introducing resistance (R) alleles from Solanum demissum (reviewed by Gebhardt and Valkonen, 2001) and more recently from Solanum bulbocastanum (Helgeson et al., 1998; Song et al., 2003) two wild potato species indigenous to Mexico. Moreover, some attempts have been done to exploit host potato resistance in order to make a more rationale use of fungicide in a sense to determine the optimal number and timing of

(23)

fungicide applications using different approaches (Fry, 1977; Fry, 1978; Fry and Shtienberg, 1990; Kankwatsa et al., 2002; Kirk et al., 2001; Kirk et al., 2005; Shtienberg et al., 1989). In the integration of host resistance and rationale use of fungicides for late blight control, it is important to consider other factors such as weather and socio-economic conditions for a specific location (Crissman et al., 1998; Ortiz et al., 2004). From the epidemiological point of view it is very important to carry out a precise and accurate assessment of host potato resistance and to know the cultivar performance under different agroecological growing conditions to develop and implement environmentally and economically effective strategies aimed at controling late blight. First steps in that direction have been taken (Andrivon et al., 2006; Hansen et al., 2005; Nærstad et al., 2007; Yuen and Forbes, 2009).

For the potato growers of the developing countries, the majority of whom have limited economic resources, the most promising alternative at the moment for late blight management is the use of varieties with elevated levels of resistance and judicious use of low-cost fungicides, such as mancozeb (Grünwald et al., 2002). Nevertheless, the determination of the timing, number and frequency of applications of fungicides in a specific locality, and with a specific cultivar, is a difficult task since it requires many field experiments due to the great variation in weather and available levels of host resistance among regions (Ortiz et al., 2004). Given these circumstances, computerized models for disease simulation can be effective tools in the evaluation of strategies for disease control as they allow preliminary assessment of a number of scenarios involving many variables, including weather conditions, level of host resistance and aggressiveness/virulence of the local pathogen population (Andrade-Piedra et al., 2005a). Disease simulation can thus reduce research costs, since only the most promising strategies would be evaluated in the field (Shtienberg et al., 1989; Shtienberg and Fry, 1990).

Initially, the LATEBLIGHT simulation was used to investigate the effect of rate-reducing resistance on the performance of the protectant fungicide chlorothalonil sprayed at fixed intervals (Bruhn and Fry, 1981). Later, an improved version of LATEBLIGHT (version LB1990) was used to examine several strategies involving the fungicides metalaxyl and chlorothalonil for late blight control and delay of metalaxyl resistance development in pathogen populations (Doster et al., 1990). Details of the model components and descriptions of how simulations are conducted have been described (Andrade-Piedra et al., 2005b; Bruhn et al., 1980; Bruhn and Fry, 1981; Doster et al., 1990; Fry et al., 1991). The most recent version of the

(24)

simulation model LATEBLIGHT, LB2004, was developed and validated by Andrade-Piedra et al. (2005b). The epidemiological parameters were measured in three Peruvian varieties infected with isolates of the ^EC-1 clonal lineage of P. infestans, which is dominant in Peru and the Northern Andes (Fry et al., 2009). In that study, the authors only simulated disease in nontreated plots (Andrade-Piedra et al., 2005b). The fungicide sub-model published by Bruhn and Fry (1982a,b) did not give satisfactory results with the data from Peru (unpublished data). The LB2004 version was subsequently used successfully with data from locations worldwide, but again without fungicide applications (Andrade-Piedra et al., 2005c).

Given the need for more effective management strategies against late blight, it was proposed that LATEBLIGHT simulation model version LB2004

could be a useful tool for initial evaluation of disease management scenarios in Nicaragua. However, it was not known if the model would work under Nicaraguan conditions either because of the epidemiological parameters used by Andrade-Piedra et al. (2005a) would not be appropriate for the Nicaraguan cultivars and pathogen population, or for other undetermined reasons. It was also proposed to evaluate the appropriateness of a modified version of the fungicide sub-model. One problem associated with the management of late blight with host resistance in developing countries is the lack of a system for quantifying host resistance. Recently, it was proposed a simple scale for quantifying resistance (susceptibility) that is putatively robust across locations (Yuen and Forbes, 2009).

(25)

3 Aims of this study

 To assess the genotypic and phenotypic variation in isolates of P.

infestans collected in potato and tomato growing areas of northern Nicaragua. In addition, this study also aimed to compare the isolates of P. infestans from potato and tomato to determine whether there was differentiation between these two groups of isolates at the genotypic and phenotypic level (Paper I).

 To establish whether there were differences in aggressiveness among potato and tomato isolates of P. infestans; to determine whether the P. infestans population from Nicaragua is formed exclusively by a single clonal lineage (Paper II).

 To evaluate the appropriateness of the LATEBLIGHT simulation model for use in Nicaragua. The specific objectives of this field study were: i) to assess the adequacy of the LATEBLIGHT simulation model (version LB2004) for disease management scenario testing under Nicaraguan conditions; ii) to quantify the degree of susceptibility to P. infestans in three potato cultivars grown in Nicaragua; and iii) to compare three application intervals of the contact fungicide, chlorothalonil for disease management with these cultivars. (Paper III).

 To determine the theoretical relationship between the intensity of fungicide use and level of host plant resistance. An additional objective was to determine an optimum fungicide treatment strategy (timing and number of fungicide applications needed) to control late blight in accordance with the levels of host plant resistance of the cultivars used in this study and the prevailing geographical weather conditions of the experimental sites (Paper IV).

(26)

(27)

4 Materials and Methods

4.1 Sampling and isolation of Phytophthora infestans (Paper I and II)

Leaflets of potato and tomato with a single late blight lesion were collected from commercial production and experimental fields in three departments (Estelí, Jinotega and Matagalpa) in northern Nicaragua from July 2007 to January 2010 (Figure 2). In each department five to seven locations were sampled, taking a different number of samples from each location (Table 1).

Thirteen tomato and forty-three potato fields (56 fields in total) located in eighteen sites in three northern departments of Nicaragua were sampled.

Details about the sampling sites and the number of isolates collected per department and particular location are shown in Table 1. The infected leaflets were washed with distilled water and dried with filter paper.

Thereafter, they were individually placed abaxial side up in a sealed Petri dish containing a layer of 1.5% water agar and incubated at 18°C to promote sporulation. When sporulation was observed, the mycelia with sporangia were transferred to a pea agar medium (Flier et al., 2003), amended with antibiotics (0.2 g ampicillin and 10 mg pimaricin L^-1) and incubated at 18°C in darkness for a week. Plugs of agar with growing hyphal tips were cut from the colony margins and transferred to Petri dishes with pea agar medium without antibiotics and incubated at 18°C for growth and sporulation. Axenic isolates were maintained on pea agar medium without antibiotics and transferred monthly to fresh medium. Of the sampled material, fifty-four blighted potato and tomato leaflets were preserved as dried material for DNA extraction.

(28)

Figure 2. Samples of single lesion potato and tomato leaflets infected with Phytophthora infestans were collected from three northern departments of Nicaragua from 2007 to 2010. A total of 248 isolates of P. infestans were obtained.

4.2 DNA extraction

Two approaches were used to extract ^DNA for mitochondrial haplotyping and microsatellite (SSR) analysis depending on whether the sample was stored as lyophilized mycelium or as dried leaflets. Individual pieces of lyophilized mycelium were placed in a 2-mL polypropylene vial containing six glass beads and homogenized in a FastPrep preparation shaker (Precellys 24, Bertin Technologies). DNA was extracted following the protocol provided with the Wizard^® Genomic ^DNA purification kit: protocol for plant tissue (Promega) for isolating genomic DNA from plant tissue. Dried leaflets of potato and tomato infected with P. infestans were homogenized for ^DNA extraction as described for lyophilized mycelium. ^DNA from dried leaflets was extracted using a cetyltrimethylammonium bromide (CTAB) procedure (Gardes and Bruns, 1993), with exception that 3% CTAB was used.

(29)

4.3 Genotypic characterization

Mitochondrial DNA (mtDNA) haplotyping was carried out using a method described earlier (Griffith and Shaw, 1998), with slight modifications. The annealing temperature was increased to 63°C and the primer pairs P2 and P4 were used at a concentration of 0.4 µM. Microsatellite analysis was carried out as described in Papers I and II. For genotypic characterization, 204 isolates of P. infestans from Nicaragua were used.

4.4 Phenotypic characterization

Mating type determination was done using tester isolates of known mating type (A1 or A2) and the unknown Nicaraguan isolates. Mycelial plugs (0.5 cm diameter) of each were placed in petri dishes containing rye pea agar (Lehtinen et al., 2008) at 20°C in the dark. Cultures were examined for oospore formation in the zone of interaction. In the mating type assays 248 Nicaraguan isolates of P. infestans were used.

The fungicide sensitivity of the isolates to metalaxyl-M and propamocarb hydrochloride (propamocarb-HCl) and the virulence testing was done using the floating leaf disc method (Lehtinen et al., 2008; Sozzi et al., 1992). For virulence testing, a procedure described earlier (Lehtinen et al., 2008) was followed. Only potato plants with resistance genes were used for virulence tests, since no tomato differentials were available. Mean number of virulence factors per isolate (Ci) and race (Cp) was calculated (Andrivon, 1994). The Ci and Cp were separately calculated for the potato and tomato isolates. Moreover, to detect differences among potato and tomato isolates, a t-test procedure for the Ci and Cp values was performed.

For aggressiveness determination, sixteen isolates from potato and fifteen isolates from tomato were used in cross-inoculation assays, .i.e., potato leaflets were individually inoculated with potato and tomato isolates and the same was done with tomato leaflets. The inoculum was prepared directly from artificially infected potato and tomato leaflets. Each isolate-host combination was repeated five times (one leaflet of potato or tomato per Petri dish). Potato or tomato leaflets were placed abaxial face up on the lids of inverted Petri dishes lined with 1.5% water agar in the base and inoculated with a 20 µL droplet of sporangial suspension adjusted to 2 x 10⁴ sporangia mL^-1 of the appropriate test isolate (potato or tomato). Thereafter, the inoculated leaflets were incubated at 16°C and 16 h day length. The incubation period (IP), latency period (LP), lesion area (LA), lesion growth rate (LGR), spore production (^SP), sporulating area (^SA) and sporulation rate (SR) were determined as described elsewhere (Andrade-Piedra et al., 2005a;

(30)

Mizubuti and Fry, 1998; Suassuna et al., 2004). An aggressiveness index (A_i) for each isolate-host combination was calculated using the following equation: ^Ai = ln (^LA × ^SP × 1/^LP) (Montarry et al., 2007; Montarry et al., 2008).

4.5 LATEBLIGHT simulation model version LB2004 (Paper III and IV)

4.5.1 Assessing the adequacy of the simulation model LATEBLIGHT under Nicaraguan conditions (Paper III)

Five field experiments were implemented in two potato growing regions in northern Nicaragua and included two susceptible (Cal White and Granola) and one resistant (Jacqueline Lee) cultivars, the latter of which was recently introduced to Nicaragua. All three had a vegetative cycle lasting between 90 and 110 days under Nicaraguan conditions. Cal White and Granola were both known to be susceptible to P. infestans, while Jacqueline Lee was considered to be resistant to the US-8 strain in the United States (Douches et al., 2001; United States Potato Board, 2007). However, quantitative information on the level of resistance was not known for any of the cultivars. Each potato cultivar was planted separately. Fertilizers and non- experimental pesticide sprays were applied in accordance with the grower practices at each of the three locations.

The fungicide treatments consisted of three application intervals of the fungicide chlorothalonil (Knight 72 SC, 720 g a.i./L). Fungicide applications were initiated for a particular cultivar-location combination when percent emergence was at 50%. The first date of fungicide application was considered as the date of crop emergence and was used in the simulations. After the first fungicide application, plots were sprayed every 4, 7 or 14 days, depending on treatment. The fungicide chlorothalonil (Knight 72 SC) was applied at the recommended rate of 1.5 L/ha and ensuring a concentration of 2.52 g a.i./L of water. Plots without fungicide application were left as nontreated controls. Percent disease severity was estimated visually once a week starting 1 to 2 days after 50% plant emergence using a late blight standard area diagram (Anonymous, 1947), which had been modified by Fry (1977). Severity values from each epidemic (year-location- cultivar-treatment combination) were converted to the area under the disease progress curve (AUDPC) using the midpoint method (Campbell and Madden, 1990), and then to the relative AUDPC (RAUDPC) as described

(31)

earlier (Fry, 1978). To evaluate the level of susceptibility of the three cultivars to P. infestans, the RAUDPC was converted to susceptibility scale values as described by Yuen and Forbes (2009).

Simulations were performed with the LB2004 version of model

LATEBLIGHT (Andrade-Piedra et al., 2005a). Initially, the parameters derived by Andrade-Piedra et al. (2005a) for the susceptible Peruvian cultivar Tomasa were evaluated in simulations with the epidemic data from Nicaragua for use with the two susceptible cultivars Cal White and Granola, but the results were not satisfactory. Therefore, modifications were made in the lesion growth rate (LGR) and the sporulation rate (SR) of the parameters from Tomasa. The modifications were made by visually fitting simulated and observed disease progress curves. The day of initiation of the epidemic and the number of initial lesions were determined by the method used by Andrade-Piedra et al. (2005b). It was evident from the disease severity data that cultivar Jacqueline Lee was protected by a major R gene and primarily had a hypersensitive resistance reaction. For this reason it was eliminated from the simulation process.

The original fungicide efficacy sub-model for LATEBLIGHT was developed and validated by Bruhn and Fry (Bruhn and Fry, 1981; Bruhn and Fry, 1982a; Bruhn and Fry, 1982b) specifically for chlorothalonil. In that model an average fungicide effect was calculated from the individual effects of the fungicide residues distributed among four levels in the canopy according to a gamma distribution (Bruhn and Fry, 1982a; Bruhn and Fry, 1982b).

In the present study, a simplified version in which there is no longer an effect of canopy level was employed. For our model, fungicide was assumed to be applied evenly on foliage. We believe this is justified in developing countries because we have observed that farmers using backpack sprayers tend to spray around the plant to achieve even coverage. The average level of deposition for foliage was estimated based on the concentration of the fungicide in the spray solution and a residue factor that indicated the amount of water that remains on the potato leaf surface when foliage is sprayed until run-off. A value of 0.0068 cm³ water/cm² leaf was used for the residue factor (van Haren and Jansen, 1999). Details about the fungicide submodel structure are found in Paper III.

The adequacy of the model was investigated in several ways. Firstly, observed and simulated disease progress curves were compared graphically to evaluate the fit between observed and predicted data of disease progress with and without fungicides, and the overall pattern of disease severity relative to increasing fungicide application. Secondly, the efficacy of the

(32)

fungicide submodel was further evaluated by examining the number of cases when simulation falsely predicted control (false positive) or falsely predicted the absence of control (false negative). Control was arbitrarily defined to occur when final disease severity did not surpass 20% (Kromann et al., 2009). Thirdly, the simulator was also evaluated based on the deviations between observed and predicted ^AUDPC values. Deviations were compared to an envelope of acceptance test (EAT), the boundaries for which are calculated from the error of the observed values (Mitchell, 1997). The relationship between fungicide application and disease development measured by the AUDPC was explored with regression analysis, using the

REG procedure of SAS (version 9.1; SAS Institute, Cary, NC). The quadratic regression equation provided the best fit for the relationship between the

AUDPC and the fungicide spray interval. Cultivars were compared for

RAUDPC values within each location-year combination using the least significant difference (LSD) test following an analysis of variance using PROC ANOVA of SAS v.9.1 (SAS Institute Inc., 2004). Normality of distributions and homogeneity of variances of experimental errors were tested as described by Quinn and Keough (2009).

4.5.2 Epidemiological significance of the quantitative relationship between host resistance and fungicide usage (Paper IV)

Simulation and field experiments were carried out with the following objectives: i) to determine the theoretical relationship between fungicide intensity and level of host plant resistance; and ii) to determine an optimum fungicide treatment strategy (timing and number of fungicide applications needed) to control late blight in accordance with the levels of host plant resistance of the cultivars and the prevailing geographical weather conditions of the experimental sites.

Nine levels of resistance were used with the simulation model

LATEBLIGHT (version ^LB2004, Andrade-Piedra et al., 2005b) to produce data points for assessment. Each resistance level represented a synthetic cultivar and was developed by altering four epidemiological parameters: latency period (^LP), lesion growth rate (^LGR) sporulation rate (^SR) and infection efficiency (IE). During 2010, two simultaneous field trials were conducted at the Experimental Station Santa Catalina, Quito, Ecuador (International Potato Center, ^CIP, by its acronym in Spanish). In each trial, 12 potato cultivars with different degree of susceptibility to P. infestans were evaluated (Table 3). In both experiments, potato plants were grown in 3 m long × 4

(33)

m wide plots (four rows per plot and 10 plants per row, planted at 0.3 m spacing in the row). Each plot was separated from each other by a strip of 1 m oat. In the first field experiment, 12 treatments (12 potato genotypes) with three replicates were evaluated for their susceptibility to local population of P. infestans. In the second field experiment (hereafter referred to as “Trial II”), 72 treatments, resulting from the combination of six fungicide spray regimes and 12 potato genotypes were evaluated. More details on simulations and field experiments are described in Paper IV.

(34)

(35)

5 Results and Discussion

5.1 Sampling and isolation of Phytophthora infestans (Paper I and II)

Of the sampled isolates, 84% (209) isolates were isolated from blighted potato leaflets and 16% (39 isolates) were isolated from blighted tomato leaflets and fruits. All of the 248 collected isolates were tested for mating type, a subset of 132 isolates were used for microsatellite analysis and mtDNA haplotyping. Ninety-eight isolates (82 from potato and 16 from tomato) were used for fungicide sensitivity and used in virulence tests.

Isolates for genotypic and phenotypic analyses were collected from 11 and 12 locations respectively (Table 1).

5.2 Genotypic and phenotypic characterization of Phytophthora infestans population from Nicaragua (Paper I and II)

5.2.1 Genotypic characterization

In the first study aimed to assess the genotypic diversity of Nicaraguan population of P. infestans, SSR genotyping using set of seven primers (4B,

G11, ^Pi16, ^Pi70, ^D13, ^Pi63 and ^Pi04) revealed no polymorphism in 121 out of 132 isolates of P. infestans from Nicaragua. The only exception to this were two rare genotypes that showed one-step difference at the loci Pi16 and

G11, respectively when compared to the commonly found genotype.

Variations at the loci Pi16 and G11 were found in one and ten potato isolates respectively, representing 0.7% for the locus Pi16 and 7.6% for the locus G11

of the total isolates tested. These eleven potato isolates were collected from three different locations (El Arenal, Miraflor and Tisey; Table 1).

(36)

Otherwise, all the other isolates from potato and tomato belonged to a single multilocus genotype hereafter referred to as NI-1 genotype. This dominant genotype was heterozygous for almost all the analyzed loci (^4B – 205/213; G11 – 132/156; D13 – 98/108; Pi63 – 148/157; and Pi04 – 166/170), except for Pi16 (176/176) and Pi70 (192/192) loci, which were found to be homozygous. Minor variants of this genotype were found with 176/174 at Pi16 (1 isolate) and 132/154 at G11 (10 isolates). Mitochondrial

DNA (mtDNA) haplotyping revealed that all 132 isolates tested had the Ia haplotype. No evidence was found of population differentiation among potato and tomato isolates of P. infestans based on the SSR fingerprinting patterns and mtDNA haplotyping (Paper I).

In a second study, 72 isolates of P. infestans (53 from potato and 19 from tomato) from Nicaragua were further genotypically characterized using the same abovementioned SSR markers and mtDNA haplotyping. Five SSR

multilocus genotypes among 72 isolates of P. infestans from Nicaragua were detected and all 72 isolates sampled from potato and tomato fields were of the Ia mtDNA haplotype and A2 mating type. The most predominant was the genotype NI-1, found in 63 out of 72 isolates and reaching a frequency of 87.5%. The N-1 genotype was common to 46 potato isolates and 17 tomato isolates. The frequency of the remaining four genotypes was very low (Figure 3). Variation in tomato isolates was found only in two isolates at loci 4B and Pi16 and in both cases they shared the same allele sizes with two potato isolates. In general, two kinds of variations were detected, namely, from heterozygosity to homozygosity at loci 4B and G11 and from homozygosity to heterozygosity at locus Pi16. The common trait of the five identified genotypes is that they belonged to the ^A2 mating type and had the Ia mtDNA haplotype. The 4B, G11 and Pi16 loci were the most variable loci, as they showed differences among tested isolates of P. infestans (Paper II).

(37)

Table 1. Origin, mating type, mitochondrial DNA haplotype and SSR fingerprinting pattern of Phytophthora infestans isolates collected from 2007 to 2010 in Northern Nicaragua.

Department Location^a Crop N-of-I^b Mating type Haplotype^c SSR^d

El Jobo^P Potato 10 A2 nd^e nd

La Laguna^G,P Potato 21 A2 Ia (13) M^f (13)

Estelí La Tejera^G,P Potato 9 A2 Ia (3) M (3)

Miraflor^G,P Potato 39 A2 Ia (37) M (34), V^g (3)

Sesteo^P Potato 23 A2 nd nd

Tisey^G,P Potato 34 A2 Ia (22) M (20), V (2)

Sub-total 6 136 136 75 75

Chagüite Grande Tomato 12 A2 nd nd

El Canal^G Tomato 7 A2 Ia (7) M (7)

El Mojón^P Potato 3 A2 nd nd

Jinotega El Mojón^G Tomato 1 A2 Ia (1) M (1)

Las Colinas^P Tomato 4 A2 nd nd

La Galia^P Potato 10 A2 nd nd

La Parranda^G Potato 5 A2 Ia (5) M (5)

Tomatoya^P Tomato 5 A2 nd nd

Subtotal 7 47 47 13 13

Aranjuez^G Potato 1 A2 Ia (1) M (1)

El Arenal^G Potato 17 A2 Ia (17) M (11), V (6)

Matagalpa La Fundadora^G Potato 29 A2 Ia (18) M (18)

La Fundadora^G,P Tomato 10 A2 Ia (3) M (3)

Sitio Viejo^G Potato 5 A2 Ia (5) M (5)

Yucul^P Potato 3 A2 nd nd

Sub-total 5 65 65 44 44

Total 18 248 248 132 132

aIsolates collected from locations marked with the letters G, P and GP, were used for genotypic (G) and phenotypic (P) analyses. In some cases the isolates were collected from the same location for both analyses (GP).

bNumber of isolates collected from 2007 to 2010 in the main potato growing areas of northern Nicaragua.

cMitochondrial DNA haplotype. In parenthesis is indicated the number of isolates that were tested.

dSSR = Simple sequence repeats (also known as microsatellites).

end = not determined or not included in the analysis.

fM = monomorphic for SSR markers. In parenthesis is indicated the number of isolates that were included in the analysis.

gV = variants, it means those isolates that showed one-step difference at loci G11 and Pi16. For instance in location Miraflor, 2 isolates had a one-step difference at locus G11 and 1 isolate showed one-step difference at locus Pi16. On the other hand, in other locations such as Tisey and El Arenal variation was observed only at locus G11.

Overall, 204 isolates of P. infestans (165 from potato and 39 from tomato) were analyzed using SSR markers and mtDNA haplotype determination in both studies (Table 2). It has been hypothesized that the first population of P. infestans present in Nicaragua belonged to the “old” single clonal lineage (US-1 genotype) of the A1 mating type and Ib mtDNA haplotype (Fry and Goodwin, 1997). This genotype probably arrived to Nicaragua in the early 1900s with a shipment of potato for consumption from United States. The Ia and IIb mtDNA haplotypes have been found in herbarium specimens from Nicaragua dating from 1954 and 1956 respectively (May and Ristaino, 2004). Our data suggests, however, that P. infestans populations have experienced a major shift since its first appearance in Nicaraguan potato fields.

(38)

Table 2. Simple sequence repeat (SSRs) multilocus genotypes detected in Phytophthora infestans isolates from Nicaragua collected from July 2007 to January 2010.

Gt^(a) NoI^(b) Host Allele sizes^(c) detected with seven SSR loci 4B G11 Pi16 Pi70 D13 Pi63 Pi04 NI-1 195 P/T^(d) 205 132 176 192 98 148 166 213 156 176 192 108 157 170

NI-2 2 P/T 205 132 176 192 98 148 166 213 156 180 192 108 157 170

NI-3 2 P 205 156 176 192 98 148 166

213 156 176 192 108 157 170

NI-4 2 P/T 213 132 176 192 98 148 166 213 156 176 192 108 157 170

NI-5 3 P 213 156 176 192 98 148 166

213 156 176 192 108 157 170 Total 204

(a)Gt: Genotypes found using seven SSR markers. The NI-1 genotype was the most predominant. The frequency of the other genotypes was very low.

(b)NoI: Number of isolates in a given genotype. The NI-1 genotype was common to 158 potato isolates and 37 tomato isolates, whereas 1 potato and 1 tomato isolate shared the same allele sizes and were grouped in the NI-2 and NI-4 genotypes.

(c)Allele sizes in bold are indicating where the variation was found. Allele sizes were adjusted to the sizes obtained by Lees et al. (2006).

(d)P/T: Potato or tomato host.

Genotypic diversity within populations of P. infestans from Nicaragua was expected due to the fact that potato seed is imported from the Netherlands, Canada, United States, and Guatemala. Contrary to this initial hypothesis, the P. infestans population from Nicaragua seems to belong to a single clonal lineage having the A2 mating type and the Ia mtDNA haplotype. Our results indicate that the Nicaraguan clonal lineage of P. infestans does not originate from seed imported from the Netherlands or other European sources, since the allele with the size 132 bp found at the locus G11 has not been recorded in European populations (D. Cooke, personal communication). This clonal lineage is not US-8, either, since that clonal lineage has a different genotype at these microsatellite loci (D. Cooke, personal communication).

(39)

0 10 20 30 40 50 60 70 80 90 100

NI-1 NI-2 NI-3 NI-4 NI-5

Percentage of isolates

Genotypes of Phytophthora infestans Potato isolates Tomato isolates

Figure 3. Genotypes of Phytophthora infestans detected using simple sequence repeat (SSRs) markers and the percentage of potato (n=165) and tomato (n=39) isolates found in each genotype.

The allele size 132 at G11 has been found in a P. infestans strain from Mexico (www.euroblight.net) and has been recorded from A1 tomato isolates from the United States such as US-11 and US-12 (D. Cooke, personal communication), suggesting a New World origin of the Nicaraguan population. In studies carried out in Venezuela using 4B and G11 SSR

markers (Briceño et al., 2009) and Colombia using 4B and D13 markers (Vargas et al., 2009), a similar low genotypic diversity was found among the tested P. infestans isolates. In Central America, the mating type reported here for the Nicaraguan population of P. infestans is the opposite of that reported from neighboring countries. Transfer of agricultural products occurs over the borders of Nicaragua, Costa Rica and Honduras and one

Insights into population structure and epidemiology of Phytophthora infestans from Nicaragua